Modern approach to diagnosis in patients with suspected deep vein thrombosis.

Deep vein thrombosis is a relatively common disease, amenable to therapy but with a potentially fatal outcome if untreated. The diagnosis can be made in most patients with the non-invasive imaging procedure ultrasonography, but limitations exist. As with all tests, there is a potential for false-positive and false-negative results. The latter are especially an issue for calf vein thrombi, and this in part has led to the concept of serial testing of the proximal venous system rather than imaging the calf. The premise of the repeat (serial) test is that only thrombi that extend to the proximal system are clinically relevant and such thrombi will be detected on the repeat test. However, despite the safety of the serial-testing concept, it is inconvenient and expensive. In the last few years, the diagnostic process has been improved by the validation of a clinical model that accurately categorizes patients as having low, moderate or high probability of deep vein thrombosis. Among the improvements this provides is the elimination of serial testing if the ultrasound is normal and the clinical probability low. The fibrin degradation product D-dimer has been demonstrated to have a high negative predictive value and has also proven useful in diagnostic algorithms. The combination of the D-dimer with clinical model assessment will enable diagnostic-testing strategies that are more safe, effective and convenient for patients.